My favorites | Sign in
Project Home Wiki Issues Source
New issue   Search
  Advanced search   Search tips   Subscriptions
Issue 74: Revisiting voiD discovery
2 people starred this issue and may be notified of changes. Back to list
Status:  Fixed
Owner:  Michael.Hausenblas
Closed:  May 2011

Blocked on:
issue 85

Sign in to add a comment
Project Member Reported by, Sep 15, 2010
Semantic Sitemaps are dead. We should no longer promote them, or rely on them for voiD discovery.

What else could we do for discovering the voiD description of a domain?

Link in robots.txt? Use .well-known? Use a pseudo-canonical http://yourdomain/void.ttl location?
Sep 30, 2010
Project Member #1
(No comment was entered for this change.)
Labels: Milestone-Release2.0
Sep 30, 2010
Project Member #2
Per today's call, let's do it in Release 2.

We won't do the pseudo-canonical location approach.
Oct 29, 2010
Project Member #3 Michael.Hausenblas
As per discussion today, resolved to outsource it and the register with IETF; changed ownership as well
Owner: Michael.Hausenblas
Dec 8, 2010
Project Member #4
I think in detail we should do this, in order:

1. Remove Section 7.2 from the Guide

2. Explain in the beginning of Section 7 that previous versions of voiD had a discovery mechanism based on robots.txt and Semantic Sitemaps, which was deprecated because it was not widely adopted. Include a link to

3. Explain in the beginning of Section 7 that a new discovery mechanism will be defined in the future in a separate document, linking to

4. Create a one-sentence placeholder document at

5. Get the *draft* of the Note published in W3C space

6. Write a submission to IETF for a name under .well-known.

7. Draft the autodiscovery document: Describe the .well-known mechanism, and explain that IETF approval is pending.
Dec 11, 2010
Project Member #5 Michael.Hausenblas
Done step 1-3 in r172 however, is down ATM, so need to wait till this is sorted to do the rest
Dec 13, 2010
Project Member #6 Michael.Hausenblas
Update: is back online, added - please review content

Jan 10, 2011
Project Member #7 Michael.Hausenblas
Notes that the registry is located at
Jan 19, 2011
hi - Have steps 5-7 and a submission to IETF taken place yet ? In the interim, if Im creating a VoID document, should I continue with sitemap/robots or use the .well-known mechanism ?
Jan 19, 2011
Project Member #9
Submission has not yet taken place. I expect that the entire registration process will take at least another few weeks.

My recommendation: If you already have the robots.txt/sitemap thing in place, then just leave it like that. Otherwise, publish the voiD file without autodiscovery, and wait for the process to complete.
Jan 20, 2011
Project Member #10 Michael.Hausenblas
I've implemented all the changes as discussed during today's (2011-01-20) voiD editor's call and now ready to submit to IETF.
Jan 20, 2011
Project Member #11 Michael.Hausenblas
The request has been sent [1] - review period ends in 14 days (that is, 3 Feb 2011).

Feb 1, 2011
Project Member #12 Michael.Hausenblas
As per today's (2011-02-01) editors call we decided that we'll publish the VoID guide as SWIG Note and then come back to IETF to register it (ETA: mid/end of Feb 2011)
Feb 3, 2011
Project Member #13
I took an action to draft a replacement for the placeholder autodiscovery section 7.2, along with clarifications regarding the relationship between hostnames and datasets. Proposed text is below.

Replace Section 7.2, “Discovery with well-known URI”:

<p><a href="">RFC 5758</a> [<a href="#ref-RFC5758">RFC5758</a>] defines a mechanism for reserving “well-known” URIs on any web server.</p>

<p>The URI <code>/.well-known/void</code> on any web server is reserved for a VoID description of any datasets hosted on that server. For example, on the host <code></code>, this URI would be <code></code>.</p>

<p>This URI may be an HTTP redirect to the location of the actual VoID file. The most appropriate HTTP redirect code is 302. Clients accessing this well-known URI MUST handle HTTP redirects.</p>

<p>The VoID file accessible via the well-known URI should contain descriptions of all RDF datasets hosted on the server. This includes any datasets that have resolvable URIs, a SPARQL endpoint, a data dump, or any other access mechanism (see <a href="#access">Section 3</a>) whose URI is on the server's hostname.</p>

<p>Any VoID file accessible via the well-known URI should follow the guidelines set out in <a href="">Section 6.2, <em>Publishing a voiD file alongside a dataset</em></a>.</p>

In addition, add this to the end of Section 6.2, “Publishing a voiD file alongside a dataset”:

<p class="note">In cases where multiple different <code>void:Dataset</code>s are published on the same website, the easiest option is usually to create a single <code>void:DatasetDescription</code> document that describes all of them.</p>
Feb 9, 2011
Project Member #14 Michael.Hausenblas
Implemented changes re auto-discovery in VoID guide as of comment #13, see r185
Status: Started
Feb 15, 2011
Project Member #15
FYI: I tweaked the formatting of the text in the IETF registration box in r186.
Feb 17, 2011
Project Member #16 Michael.Hausenblas
(No comment was entered for this change.)
Blockedon: 85
May 12, 2011
Project Member #17 Michael.Hausenblas
Yey! IANA registration completed - see announcement at [1]. Guess it's time to close the last v2 issue then ;) *party*
Status: Fixed
May 12, 2011
Project Member #18 Michael.Hausenblas
... and I sure did forget the pointer to the announcement ... here it comes:
Sign in to add a comment

Powered by Google Project Hosting